BR122020009714B1 - METHOD OF DECODING A VIDEO, APPARATUS OF ENCODING A VIDEO, AND NON-TRANSIVE COMPUTER-READABLE MEANS FOR STORING DATA ASSOCIATED WITH A VIDEO - Google Patents

Abstract

methods for encoding and decoding the entropy of a video are provided. the entropy decoding method includes obtaining a transform unit significant coefficient indicator that indicates whether a non-zero transform coefficient exists in the transform unit, from a bit stream (bit stream), determining a context model to arithmetically decode the transformation unit significant coefficient indicator based on the transformation unit depth of transformation and arithmetically decode the transformation unit significant coefficient indicator based on the given context model

Description

Technical Field

[0001] One or more embodiments of the present invention relate to video encoding and decoding, and more particularly to a method and apparatus for encoding and entropy encoding information related to a transformation unit.

Background of the Technique

[0002] According to image compression methods such as advanced video encoding (AVC) of MPEG-1, MPEG-2, MPEG-4 or MPEG-4 H.264/MPEG-4, an image is divided in blocks having a predetermined size, and then the residual data of the blocks is obtained by Interprediction, or intraprediction. Residual data is compressed by transformation, quantization, digitization, run-length encoding, and entropy encoding. In entropy encoding, a syntax element such as a transform coefficient or a prediction mode is entropy encoded to produce a bit stream. A decoder parses and extracts syntax elements from a bitstream, and reconstructs an image based on the extracted syntax elements.

Disclosure of Invention Technical Problem

[0003] One or more embodiments of the present invention include an entropy encoding method and apparatus, and an entropy decoding method and apparatus for selecting a context model used to entropy encode and decode a syntax element related to a transformation unit which is a unit of data used to transform a coding unit, based on a transformation depth, indicating a hierarchical division relationship between the coding unit and the transformation unit.

Technical Solution

[0004] A context model for arithmetically decoding a transformation unit significant coefficient indicator is determined based on the depth of the transformation, indicating the number of times the coding unit is divided to determine the transformation unit included in the transformation unit. encoding, and the transform unit significant coefficient indicator is arithmetically decoded based on the given context model.

Advantageous Effects

[0005] According to embodiments of the present invention, by selecting a context model based on a depth of transformation, a condition to select the context model can be simplified and the operation for encoding and decoding by entropy can also be simplified.

Brief Description of Drawings

[0006] FIG. 1 is a block diagram of a video encoding apparatus in accordance with an embodiment of the present invention.

[0007] FIG. 2 is a block diagram of a video decoding apparatus in accordance with an embodiment of the present invention.

[0008] FIG. 3 is a diagram for describing a concept of encoding units in accordance with an embodiment of the present invention.

[0009] FIG. 4 is a block diagram of a video encoder based on encoding units having a hierarchical structure, in accordance with an embodiment of the present invention.

[00010] FIG. 5 is a block diagram of a video decoder based on encoding units having a hierarchical structure, in accordance with an embodiment of the present invention.

[00011] FIG. 6 is a diagram illustrating deeper encoding units according to depths, and partitions, in accordance with an embodiment of the present invention.

[00012] FIG. 7 is a diagram for describing a relationship between an encoding unit and transforming units, in accordance with an embodiment of the present invention.

[00013] FIG. 8 is a diagram describing coding information from coding units corresponding to an encoded depth, in accordance with an embodiment of the present invention.

[00014] FIG. 9 is a diagram of deeper coding units according to depth, in accordance with an embodiment of the present invention.

[00015] FIGS. 10 to 12 are diagrams for describing a relationship between coding units, prediction units, and frequency transformation units, in accordance with an embodiment of the present invention.

[00016] FIG. 13 is a diagram for describing a relationship between a coding unit, a prediction unit, and a transformation unit, according to the coding mode information of Table 1.

[00017] FIG. 14 is a block diagram of an entropy encoding apparatus in accordance with an embodiment of the present invention.

[00018] FIG. 15 is a flowchart of an entropy encoding and decoding operation of a syntax element related to a transformation unit, in accordance with an embodiment of the present invention.

[00019] FIG. 16 is a diagram illustrating an encoding unit and transformation units included in the encoding unit, according to an embodiment of the present invention.

[00020] FIG. 17 is a diagram illustrating a context enhancement parameter used to determine a context model of a transform unit significant coefficient indicator of each of the transform units of FIG. 16, based on a transformation depth.

[00021] FIG. 18 is a diagram illustrating an encoding unit and a transforming unit included in the encoding unit, according to another embodiment of the present invention.

[00022] FIG. 19 is a diagram illustrating division transform indicators used to determine the structure of the transform units included in the encoding unit of FIG. 16, in accordance with an embodiment of the present invention.

[00023] FIG. 20 illustrates a transforming unit that is entropy encoded in accordance with an embodiment of the present invention;

[00024] FIG. 21 illustrates a corresponding significance map for the transformation unit of FIG. 20;

[00025] FIG. 22 illustrates the coeff_abs_level_greater1_flag indicator corresponding to the 4 x 4 transform unit of FIG. 20;

[00026] FIG. 23 illustrates the coeff_abs_level_greater2_flag indicator corresponding to the 4 x 4 transform unit of FIG. 20;

[00027] FIG. 24 illustrates the coeff_abs_level_remaining indicator corresponding to the 4 x 4 transform unit of FIG. 20;

[00028] FIG. 25 is a flowchart of a method of entropy coding a video, in accordance with an embodiment of the present invention.

[00029] FIG. 26 is a block diagram of an entropy decoding apparatus in accordance with an embodiment of the present invention;

[00030] FIG. 27 is a flowchart of a method of entropy decoding a video, in accordance with an embodiment of the present invention. Best way to carry out the Invention

[00031] According to one or more embodiments of the present invention, a method of entropy decoding a video is provided, the method includes determining a transformation unit included in an encoding unit and used to inversely transform the encoding unit; obtaining a transform unit significant coefficient indicator that indicates whether a non-zero transform coefficient exists in the transform unit, from a bit stream; if the number of times the encoding unit is divided in order to determine whether the transform unit is referred to as a transform unit transform depth, determine a context model to arithmetically decode the transform unit significant coefficient indicator, based on the transformation depth of the transformation unit; and arithmetically decode the transformation unit's significant coefficient indicator based on the given context model.

[00032] According to one or more embodiments of the present invention, an apparatus for decoding by entropy of a video, the apparatus includes an analyzer to obtain a transformation unit significant coefficient indicator that indicates whether a transformation coefficient is different from zero exists in a transform unit included in an encoding unit and used to inversely transform the encoding unit from a bit stream; a context modeler to, if the number of times the encoding unit is divided to determine whether unit transform is referred to as a transform unit depth, determine a context model to arithmetically decode the unit significant coefficient indicator transformation, based on the transformation depth of the transformation unit; and an arithmetic decoder for arithmetically decoding the transform unit significant coefficient indicator based on the given context model.

[00033] According to one or more embodiments of the present invention, a method of entropy encoding a video is provided, the method includes obtaining data from a transformed encoding unit based on a transforming unit; if the number of times the encoding unit is divided in order to determine whether the transform unit is referred to as a transform unit transform depth, determine a context model to arithmetically encode a transform unit significant coefficient indicator that indicates whether a non-zero transformation coefficient exists in the transformation unit, based on the transformation depth of the transformation unit; and arithmetically encode the transformation unit's significant coefficient indicator based on the given context model.

[00034] According to one or more embodiments of the present invention, an apparatus for encoding by entropy of a video is provided, the apparatus includes a context modeler for obtaining data from an encoding unit transformed based on a unit of transformation and, if the number of times the coding unit is divided to determine whether the transformation unit is referred to as a transformation unit depth of transformation, determine a context model to arithmetically encode a transformation unit significant coefficient indicator that indicates whether a non-zero transformation coefficient exists in the transformation unit, based on the transformation depth of the transformation unit; and an arithmetic coder for arithmetically coding the transform unit significant coefficient indicator based on the given context model. Mode for the Invention

[00035] Hereinafter, a method and apparatus for updating a parameter used in entropy encoding and decoding of dimension information of a transformation unit according to an embodiment of the present invention will be described with reference to FIGS. 1 to 13. In addition, a method of entropy encoding and decoding a syntax element obtained using the entropy encoding and decoding method of a video described with reference to FIGS. 1 to 13 will be described in detail with reference to FIGS. 14 to 27. Expressions such as "at least one of", when preceding a list of elements, modify the entire list of elements and do not modify the individual elements in the list.

[00036] FIG. 1 is a block diagram of a video encoding apparatus 100 in accordance with an embodiment of the present invention.

[00037] The video encoding apparatus 100 includes a hierarchical encoder 110 and an entropy encoder 120.

[00038] Hierarchical encoder 110 can divide a current frame to be encoded into predetermined data units to perform encoding in each of the data units. In detail, the hierarchical encoder 110 can split a current frame, based on a larger encoding unit, which is a maximum-sized encoding unit. The larger encoding unit according to an embodiment of the present invention may be a data unit having a dimension of 32 x 32, 64 x 64, 128 x 128, 256 x 256, etc., in which a unit form of dice is a square that has width and length in squares of 2 and is greater than 8.

[00039] The encoding unit according to an embodiment of the present invention can be characterized by a maximum dimension and a depth. Depth denotes the number of times the encoding unit is spatially divided from the larger encoding unit, and as the depth deepens, the deeper encoding units according to depths can be separated from the larger encoding unit for a smaller encoding unit. A greater encoding unit depth is a greater depth and a lesser encoding unit depth is a lesser depth. Since the size of an encoding unit corresponding to each depth decreases as the depth of the greater encoding unit gets deeper, an encoding unit corresponding to a greater depth may include a plurality of encoding units corresponding to lower depths.

[00040] As described above, the picture data of the current structure is divided into larger coding units according to a maximum coding unit size, and each of the larger coding units can include deeper coding units which are divided according to depth. Since the major encoding unit according to an embodiment of the present invention is divided according to depth, image data of a spatial domain included in the major encoding unit can be hierarchically classified according to depths.

[00041] The maximum depth and a maximum dimension of a coding unit, which limits the total number of times that a height and a width of the larger coding unit are hierarchically divided, can be predetermined.

[00042] Hierarchical encoder 110 encodes at least one division region obtained by dividing a region of the larger encoding unit according to depth, and determines a depth to produce finally encoded image data according to at least one region of separation. In other words, hierarchical encoder 110 determines a coded depth by encoding image data in the deepest encoding units according to the depth, according to the largest encoding unit of the current structure, and selecting a depth that has the minimal coding error. The determined coded depth and coded image data according to maximum coding units are produced for the entropy encoder 120.

[00043] The image data in the largest encoding unit is encoded based on the deepest encoding units corresponding to at least a depth equal to or less than the maximum depth, and the encoding results of the image data are compared based on each of the deeper coding units. A depth having the minimum coding error can be selected after comparing the coding errors of the deeper coding units. At least one encoded depth can be selected for each major encoding unit.

[00044] The size of the larger coding unit is divided as a coding unit is hierarchically divided according to depths and as the number of coding units increases. Furthermore, even if the encoding units correspond to the same depth into a larger encoding unit, it is determined whether to divide each of the encoding units corresponding to the same depth to a lower depth by measuring a data encoding error of each encoding unit separately. Thus, even when image data is included in a larger coding unit, the image data is divided into regions according to depth, and coding errors may differ according to regions of a larger coding unit and , thus, the encoded depths may differ according to regions of the image data. Thus, one or more coded depths can be determined in a larger coding unit, and the image data of the larger coding unit can be divided according to the coding units of at least one coded depth.

[00045] In this way, the hierarchical encoder 110 can determine the encoding units that have a tree structure included in the larger encoding unit. The 'coding units having a tree structure' according to an embodiment of the present invention, include the encoding units corresponding to a depth determined to be the encoded depth, among all the deeper encoding units included in the largest encoding unit . A coding unit having a coded depth can be determined hierarchically according to depths in the same region as the larger coding unit, and can be independently determined in different regions. Likewise, a depth encoded in a current region can be independently determined from a depth encoded in another region.

[00046] A maximum depth according to an embodiment of the present invention is an index related to the number of times a larger encoding unit is divided into smaller encoding units. A first maximum depth according to an embodiment of the present invention may denote the total number of times the larger encoding unit is divided into smaller encoding units. A second maximum depth according to an embodiment of the present invention may denote the total number of levels of depth from the largest encoding unit to the smallest encoding unit. For example, when a greater encoding unit depth is 0, a depth of one encoding unit, in which the greater encoding unit is divided once, can be set to 1, and a depth of one encoding unit, in which the largest encoding unit is divided by twice, can be set to 2. Here, if the smallest encoding unit is an encoding unit where the largest encoding unit is divided four times, five levels of depth from the 0 depths, 1, 2, 3, and 4 exist, so the first maximum depth can be set to 4, and the second maximum depth can be set to 5.

[00047] Coding and prediction transformation can be performed according to the largest coding unit. Coding and prediction transformation are also performed according to the deepest coding units according to a depth equal to or less than the maximum depth, according to the largest coding unit.

[00048] Since the number of deeper coding units increases whenever the larger coding unit is divided according to depth, coding including prediction coding and transformation is performed on all more coding units. generated as the depth increases. For convenience of description, prediction coding and transformation will now be described based on a coding unit of a current depth, in a larger coding unit.

[00049] The video encoding apparatus 100 can variously select a size or format of a data unit to encode the image data. In order to encode the image data operations such as prediction encoding, transformation, and entropy encoding are performed, and at this time, the same data unit can be used for all operations or different data units can be used for each operation.

[00050] For example, the video encoding apparatus 100 can select not only an encoding unit to encode the picture data, but also a data unit other than the encoding unit in order to perform prediction encoding on the data. of image in the encoding unit.

[00051] In order to perform prediction coding in the larger coding unit, prediction coding can be performed based on a coding unit corresponding to a coded depth, that is, based on a coding unit that is not divided into coding units corresponding to a lower depth. Hereafter, the coding unit that is no longer split and becomes a base unit for prediction coding will be referred to as a 'prediction unit'. A partition obtained by dividing the prediction unit may include a prediction unit or a data unit obtained by separating at least one of a height and a width of the prediction unit.

[00052] For example, when a 2N x 2N encoding unit (where N is a positive integer) is no longer divided and becomes a 2N x 2N prediction unit, a partition dimension can be 2N x 2N , 2NxN, N x 2N, or N x N. Examples of a partition type include symmetric partitions, which are obtained by symmetrically dividing a height or width of the prediction unit, partitions obtained by asymmetrically dividing the height or width of the unit of prediction, such as 1:n or n:1, partitions that are obtained by geometrically dividing the prediction unit, and partitions having arbitrary shapes.

[00053] A prediction mode of the prediction unit can be at least one of an intra mode, an inter mode, and a skip mode. For example, intra mode or inter mode can be performed between 2N x 2N, 2NxN, N x 2N, or N x N partition. Also, skip mode can be performed only on 2N x 2N partition. Coding is performed independently on a prediction unit in a coding unit, thus selecting a prediction mode having minimal coding error.

[00054] The video encoding apparatus 100 can also perform the transformation on the image data in an encoding unit based not only on the encoding unit for encoding the image data, but also based on a data unit that is different from the encoding unit.

[00055] In order to perform the transformation in the encoding unit, the transformation can be performed based on a data unit that has a dimension equal to or smaller than the dimension of the encoding unit. For example, the data unit for the transformation may include an intra mode data unit and an inter mode data unit.

[00056] A unit of data used as a base of the transformation is referred to as a "transformation unit". Like the coding unit, the transformation unit in the coding unit can be divided recursively into regions of smaller dimensions, so that the transformation unit can be determined individually in units of regions. Thus, the residual data in the coding unit can be divided according to the transformation unit, with the tree structure according to the transformation depths.

[00057] A transformation depth indicating the number of times the height and width of the coding unit are divided to reach the transformation unit can also be fixed in the transformation unit. For example, in a current coding unit of 2N x 2N, a transformation depth can be 0 when the dimension of a transformation unit is 2N x 2N, it can be 1 when the dimension of a transformation unit is NxN, and it can be 2 when the dimension of a transformation unit is N/2 x N/2. That is, the transformation unit having the tree structure can also be defined according to the transformation depth.

[00058] The coding information according to the coding units corresponding to a coded depth requires not only information about the coded depth, but also about the information related to prediction and transformation coding. In this way, the hierarchical encoder 110 not only determines a coded depth which has the minimum coding error, but also determines a partition type into a prediction unit, a prediction mode according to the prediction units, and a dimension of a transformation unit for transformation.

[00059] Encoding units according to a tree structure into a larger encoding unit and a method of determining a partition, according to the embodiments of the present invention, will be described in detail below with reference to FIGS. 3 to 12.

[00060] Hierarchical encoder 110 can measure a coding error of the deepest coding units according to the depths using Rate-Distortion Optimization based on Lagrange multipliers.

[00061] The entropy encoder 120 produces the image data of the larger encoding unit, which is encoded based on the at least one encoded depth determined by the hierarchical encoder 110, and on the encoding mode information according to the encoded depth , in continuous data streams. The encoded image data can be a result of encoding residual data of an image. Coding mode information according to coded depth can include coded depth information, information about partition type in prediction unit, prediction mode information, transformation unit dimension information. In particular, as will be described below, the entropy encoder 120 can entropy code a transform unit significant coefficient indicator (coded_block_flag) cbf indicating whether a transform coefficient other than 0 is not included in a transform unit, using a context model determined based on a transformation depth of the transformation unit. An operation of entropy encoding syntax elements related to a transformation unit in the entropy encoding unit 120 will be described below.

[00062] The encoded depth information can be defined using the division information according to depth, which indicates whether the encoding is performed in encoding units of a lower depth rather than the actual depth. If the current depth of the current encoding unit is the encoded depth, the image data in the current encoding unit is encoded and output, and so the information separation cannot be set to split the current encoding unit to a lower depth. Alternatively, if the current depth of the current encoding unit is not the encoded depth, encoding is performed in the smallest depth encoding unit, and therefore the division information can be set to divide the current encoding unit to get the units coding of the smallest depth.

[00063] If the current depth is not the encoded depth, encoding is performed in encoding unit which is divided into smaller depth encoding unit. Since at least one coding unit of lesser depth exists in a coding unit of the current depth, encoding is performed repeatedly in each coding unit of lesser depth, and thus encoding can be performed recursively. for encoding units that have the same depth.

[00064] Since the encoding units having a tree structure are determined for a larger encoding unit, and the information about at least one encoding mode is determined by an encoding unit of an encoded depth, the information about by Less one encoding mode can be determined for a larger encoding unit. Furthermore, an encoded depth of the image data of the larger encoding unit may be different according to locations since the image data is hierarchically divided according to depth and thus the information about the encoded depth and the Encoding mode can be set for image data.

[00065] In this way, the entropy encoder 120 can assign coding information about a corresponding coded depth and an coding mode to at least one of the coding unit, the prediction unit, and a minimum unit included in the coding unit. higher encoding.

[00066] The minimum unit according to an embodiment of the present invention is a square-shaped data unit obtained by dividing the smallest encoding unit that constitutes the lowest depth by 4. Alternatively, the minimum unit may be a unit of maximum square-shaped data that can be included in all encoding units, prediction units, partition units, and transformation units included in the largest encoding unit.

[00067] For example, coding information output through entropy encoder 120 can be classified into coding information according to coding units and coding information according to prediction units. Encoding information according to encoding units can include information about the prediction mode and the size of the partitions. The coding information according to the prediction units can include information about an estimated direction in an inter mode, about a reference picture index in an inter mode, about a motion vector, about a chroma component in a mode intra, and about an intra mode interpolation method. In addition, information about a maximum encoding unit dimension defined according to frames, slices, or GOPs, and information about the maximum depth can be inserted into a header of a bitstream.

[00068] In the video encoding apparatus 100, the deeper encoding unit may be an encoding unit obtained by dividing a height or width of an encoding unit of a greater depth, which is one layer above, by two. In other words, when the dimension of the current depth coding unit is 2N x 2N, the dimension of the smallest depth coding unit is N x N. Also, the current depth coding unit having the dimension of 2N x 2N it can include a maximum number of four lower depth coding units.

[00069] As a result, the video coding apparatus 100 can form the coding units having the tree structure by determining coding units with an optimal shape and an optimal dimension for each larger coding unit, based on the size of the larger coding unit and the maximum depth determined having characteristics of the current structure. Furthermore, since encoding can be performed in each major encoding unit using any of several prediction modes and transformations, an optimal encoding mode can be determined considering the characteristics of the multidimensional image encoding unit.

[00070] Thus, if a picture having a high resolution or a large amount of data is encoded in a conventional macroblock, a number of macroblocks per picture increases excessively. Thus, a number of pieces of compressed information generated for each macroblock increases and therefore it is difficult to transmit the compressed information and data compression efficiency decreases. However, by using the video encoding device 100, the image compression efficiency can be increased since an encoding unit is adjusted at the same time, considering the characteristics of an image while increasing a maximum dimension of a encoding unit while considering image size.

[00071] FIG. 2 is a block diagram of a video decoding apparatus 200 in accordance with an embodiment of the present invention.

[00072] The video decoding apparatus 200 includes an analyzer 210, an entropy decoder 220, and a hierarchical decoder 230. The definitions of various terms such as an encoding unit, a depth, a prediction unit, a unit transformation, and information about the various encoding modes, for various operations of the video decoding apparatus 200 are identical to those described with reference to FIG. 1 and to the video encoding apparatus 100.

[00073] Parser 210 receives a bit stream of encoded video to parse a syntax element. The entropy decoder 220 arithmetically decodes the syntax elements indicating the encoded image data based on units that have a structure by performing entropy decoding of analyzed syntax elements, and outputs the arithmetically decoded syntax elements to the hierarchical decoder 230. That is, the entropy decoder 220 performs entropy decoding of the syntax elements that are received in the form of bit sequences of 0 and 1, thus reconstructing the syntax elements.

[00074] Furthermore, the entropy decoder 220 extracts the information about a coded depth, a coding mode, the color component information, prediction mode information, etc., to the coding units having a structure of tree according to each major encoding unit, from the analyzed bit stream. The extracted information about encoded depth and encoding mode is output to hierarchical decoder 230. The image data in a bit stream is divided into the larger encoding unit so that hierarchical decoder 230 can decode the image data to each larger encoding unit.

[00075] The coded depth information and the coding mode according to the largest coding unit can be information about at least one coding unit corresponding to the coded depth, and the information about an coding mode can include information about a partition type of a corresponding encoding unit that corresponds to the encoded depth, over a prediction mode, and a dimension of a transformation unit. Furthermore, the division information according to depth can be extracted as the coded depth information.

[00076] The coded depth and coding mode information according to each larger coding unit extracted by the entropy decoder 220 is information about a coded depth and a coding mode determined to generate a minimum coding error, when an encoder, such as video encoding apparatus 100, repeatedly performs encoding for each deeper encoding unit in accordance with depths in accordance with each greater encoding unit. As a result, the video decoding apparatus 200 can reconstruct an image by decoding the image data according to an encoded depth and an encoding mode that generates the minimum encoding error.

[00077] Since the coding information about the coded depth and the coding mode can be assigned to a predetermined data unit among a corresponding coding unit, a prediction unit, and a minimum unit, the entropy decoder 220 can extract the information about encoded depth and encoding mode according to predetermined data units. When information about an encoded depth and encoding mode of a corresponding larger encoding unit is assigned to each of the predetermined data units, the predetermined data units for which the same encoded depth information and encoding mode are assigned can be inferred to be the data units included in the same larger encoding unit.

[00078] Furthermore, as will be described below, the entropy decoder 220 can entropy decode a cbf transformation unit significant coefficient indicator using a context model determined based on a transformation unit depth. An entropy decoding operation of the syntax elements referring to a transformation unit in the entropy decoder 220 will be described below.

[00079] Hierarchical decoder 230 reconstructs the current structure by decoding the picture data in each major encoding unit based on information about the encoded depth and encoding mode according to the major encoding units. In other words, the hierarchical decoder 230 can decode the encoded image data based on the extracted information about the partition type, the prediction mode, and the transformation unit for each encoding unit among the encoding units having the structure of tree included in each larger encoding unit. The decoding operation can include prediction including intra prediction and motion compensation and inverse transformation.

[00080] Hierarchical decoder 230 can perform intra prediction or motion compensation according to a partition and a prediction mode of each coding unit, based on information about the partition type and prediction mode of the prediction unit of the encoding unit according to the encoded depths.

[00081] In addition, the hierarchical decoder 230 can perform the inverse transformation according to each transformation unit in the coding unit, based on the information about the size of the coding unit's transformation unit according to the coded depths, of in order to perform the inverse transformation according to the larger encoding units.

[00082] Hierarchical decoder 230 can determine at least one encoded depth from a current larger encoding unit using division information according to depths. If the split information indicates that the image data is no longer split at the current depth, the current depth is an encoded depth. Consequently, the hierarchical decoder 230 can decode the current depth coding unit with respect to the image data of the current largest coding unit based on information about the prediction unit partition type, prediction mode, and the size of the processing unit.

[00083] In other words, the data units containing the coding information including the same division information can be collected by observing the defined coding information assigned to the predetermined data unit among the coding unit, the coding unit. prediction, and the minimum unit, and the collected data units can be considered to be a data unit to be decoded by the hierarchical decoder 230 in the same encoding mode.

[00084] The video decoding apparatus 200 can obtain information about at least one encoding unit that generates the minimum encoding error when encoding is recursively performed for each major encoding unit, and can use the information to decode the actual structure. In other words, the encoded picture data of the encoding units having the tree structure determined to be the optimal encoding units in each larger encoding unit can be decoded.

[00085] In this way, even if the image data has a high resolution and a large amount of data, the image data can be efficiently decoded and reconstructed using a dimension of one encoding unit and one encoding mode, the which are adaptively determined according to the characteristics of the image data, using information about an optimal encoding mode received from an encoder.

[00086] A method of determining encoding units having a tree structure, a prediction unit, and a transformation unit, according to an embodiment of the present invention, will now be described with reference to FIGS. from 3 to 13.

[00087] FIG. 3 is a diagram describing a concept of encoding units in accordance with an embodiment of the present invention.

[00088] A dimension of an encoding unit can be expressed in height x width, and can be 64 x 64, 32 x 32, 16 x 16, and 8 x 8. The 64 x 64 encoding unit can be split on 64x64, 64x32, 32x64, or 32x32 partitions; and a 32x32 encoding unit can be divided into 32x32, 32x16, 16x32, or 16x16 partitions; a 16x16 encoding unit can be divided into 16x16, 16x8, 8x16, or 8x8 partitions; and an 8x8 encoding unit can be divided into 8x8, 8x4, 4x8 or 4x4 partitions.

[00089] With respect to 310 video data, a resolution of 1920 x 1080, a maximum dimension of a 64 encoding unit, and a maximum depth of 2 are defined. With respect to 320 video data, a resolution of 1920 x 1080, a maximum dimension of a 64 encoding unit, and a maximum depth of 3 are defined. With respect to video data 330, a resolution of 352 x 288, a maximum dimension of a 16 encoding unit, and a maximum depth of 1 are defined. The maximum depth shown in FIG. 3 indicates a total number of divisions from a larger coding unit to a smaller coding unit.

[00090] If a resolution is high or a data amount is large, a maximum dimension of an encoding unit may be large, so as not only to increase encoding efficiency, but also to accurately reflect the characteristics of an image . As a result, the maximum size of the encoding unit of the video data 310 and 320 having the higher resolution than the video data 330 can be 64.

[00091] Since the maximum depth of the video data 310 is 2, the encoding units 315 of the video data 310 may include a larger encoding unit having a long axis dimension of 64, and the encoding units having 32 and 16 axis length dimensions as the depths are deepened to two layers, splitting the larger encoding unit twice. Meanwhile, since the maximum depth of the video data 330 is 1, the encoding units 335 of the video data 330 may include a larger encoding unit having a long axis dimension of 16, and the encoding units having a 8 long axis dimension as depths are deepened to one layer by dividing the larger encoding unit once.

[00092] Since the maximum depth of video data 320 is 3, encoding units 325 of video data 320 may include a larger encoding unit having a long axis dimension of 64 and encoding units having dimensions 32, 16, and 8 long axis as the depths are deepened to 3 layers dividing the larger encoding unit three times. As depth deepens, detailed information can be accurately expressed.

[00093] FIG. 4 is a block diagram of a video encoder 400 based on encoding units having a hierarchical structure, in accordance with an embodiment of the present invention.

[00094] An intrapredictor 410 performs intraprediction in coding units in an intra mode, with respect to a current structure 405, and a motion estimator 420 and a motion compensator 425 perform, respectively, the interestimate and the motion compensation in the units coding in an intra mode using the current 405 frame and a 495 reference frame.

[00095] The data output from the intrapredictor 410, the motion estimator 420, and the motion compensator 425 is produced as a quantized transformation coefficient through a transformer 430 and a quantizer 440. The quantized transformation coefficient is reconstructed as data in a spatial domain through an inverse quantizer 460 and an inverse transformer 470, and the reconstructed data in the spatial domain is transmitted as the reference frame 495 after being post-processed through an unblocking filter 480 and a loop filter 490. The quantized transform coefficient can be produced as a bit stream 455 through an entropy encoder 450.

[00096] The entropy coding unit 450 arithmetically encodes syntax elements referring to a transformation unit, such as a transformation unit significant coefficient indicator (cbf), indicating whether a transformation coefficient other than 0 is included in a transformation unit, a significance map that indicates a location of a transformation coefficient other than 0, a first critical value indicator (coeff_abs_level_greater1_flag) that indicates whether a transformation coefficient has a value greater than 1, a second value indicator critical (coeff_abs_level_greather2_flag) which indicates whether a transformation coefficient has a value greater than 2, and a dimension information of a transformation coefficient (coeff_abs_level_remaining) corresponding to a difference between a base level (BaseLevel) that is determined based on first critical value indicator and on the second critical value indicator and a coe efficient real transformation (abscoeff).

[00097] In order for the video encoder 400 to be applied in the video encoding apparatus 100, all the elements of the video encoder 400, i.e. the intrapredictor 410, the motion estimator 420, the motion compensator 425, the transformer 430, quantizer 440, entropy encoder 450, inverse quantizer 460, inverse transformer 470, deblocking filter 480, and loop filter 490 must perform operations based on each encoding unit between the units of encoding having a tree structure when considering the maximum depth of each larger encoding unit.

[00098] Specifically, intrapredictor 410, motion estimator 420, and motion compensator 425 determine partitions and a prediction mode of each encoding unit among encoding units having a tree structure, considering the maximum dimension and the maximum depth of a greater current coding unit, and the transformer 430 determines the size of the transformation unit in each coding unit among the coding units having a tree structure.

[00099] FIG. 5 is a block diagram of a video decoder 500 based on encoding units, in accordance with an embodiment of the present invention.

[000100] An analyzer 510 analyzes the encoded image data to be decoded and the encoding information required for decoding from a bit stream 505. The encoded image data passes through decoder 520 and inverse quantizer 530 to be produced as inversely quantified data. The entropy decoder 520 obtains the elements relating to a transformation unit from a bit stream, that is, from a transformation unit significant coefficient indicator (cbf), indicating a transformation coefficient other than 0 is included in a transformation unit, a significance map that indicates a location of a transformation coefficient other than 0, a first critical value indicator (coeff_abs_level_greater1_flag) that indicates whether a transformation coefficient has a value greater than 1, a second critical value indicator (coeff_abs_level_greather2_flag) that indicates whether a transformation coefficient has a value greater than 2, and a dimension information of a transformation coefficient (coeff_abs_level_remaining) corresponding to a difference between a base level (BaseLevel) that is determined based on the first critical value indicator and the second critical value indicator and a tra coefficient. n real information (abscoeff), and arithmetically decode the obtained syntax elements in order to reconstruct the syntax elements.

[000101] An inverse transformer 540 reconstructs the inversely quantized data to the image data in a spatial domain. An intrapredictor 550 performs intraprediction in encoding units intra-mode with respect to image data in the spatial domain, and a motion compensator 560 performs motion compensation in encoding units inter-mode using a frame of reference 585.

[000102] The spatial domain image data, which passed through the intrapredictor 550 and the motion compensator 560, can be processed as a reconstructed structure 595 after being post-processed through an unlocking filter 570 and a loop filter 580. In addition, the image data that is post-processed through the unblocking filter 570 and the loop filter 580 can be produced as the reference frame 585.

[000103] In order for the video decoder 500 to be applied to the video decoding apparatus 200, all elements of the video decoder 500, i.e. the analyzer 510, the entropy decoder 520, the inverse quantizer 530, the inverse transformer 540, intrapredictor 550, motion compensator 560, unblocking filter 570, and loop filter 580, perform operations based on encoding units having a tree structure for each larger encoding unit.

[000104] The intrapredictor 550 and the motion compensator 560 determine a partition and a prediction mode for each encoding unit having a tree structure, and the inverse transformer 540 has to determine a size of a transformation unit for each encoding unit. codification.

[000105] FIG. 6 is a diagram illustrating deeper encoding units according to depths, and partitions, in accordance with an embodiment of the present invention.

[000106] The video encoding apparatus 100 and the video encoding apparatus 200 use hierarchical encoding units in order to consider the characteristics of a picture. A maximum height, maximum width and maximum depth of the encoding units can be adaptively determined according to the characteristics of the image, or can be defined differently by a user. The dimensions of the deepest encoding units according to the depths can be determined according to the predetermined maximum dimension of the encoding unit.

[000107] In a 600 hierarchical structure of the encoding units according to an embodiment of the present invention, the maximum height and maximum width of the encoding units are each 64, and the maximum depth is 4. Since the depth deepens along a vertical axis of hierarchical structure 600, a height and width of the deeper encoding unit are each divided. In addition, a prediction unit and the partitions, which are bases for prediction encoding of each deeper encoding unit, are shown along a horizontal axis of the 600 hierarchical structure.

[000108] In other words, a coding unit 610 is a larger coding unit in the 600 hierarchical structure, where the depth is 0 and one dimension, that is, one height per width, is 64 x 64. The depth deepens along the vertical axis, and a coding unit 620 with a dimension of 32 x 32 and a depth of 1, a coding unit 630 with a dimension of 16 x 16 and a depth of 2, a coding unit 640 with a dimension of 8 x 8 and a depth of 3, and a coding unit 650 with a dimension of 4 x 4 and a depth of 4 exist. Encoding unit 650 with a dimension of 4 x 4 and depth of 4 is a smaller encoding unit.

[000109] The prediction unit and the partitions of an encoding unit are arranged along the horizontal axis according to each depth. In other words, if the encoding unit 610 with the dimension 64 x 64 and the depth of 0 is a prediction unit, the prediction unit can be divided into partitions included in the encoding unit 610, that is, a partition 610 with a dimension of 64 x 64, 612 partitions with a dimension of 64 x 32, 614 partitions with a dimension of 32 x 64, or 616 partitions with a dimension of 32 x 32.

[000110] Similarly, an encoding unit 620 prediction unit with the dimension 32 x 32 and the depth of 1 can be divided into partitions included in the encoding unit 620, i.e., a partition 620 having a dimension 32 x 32, partitions 622 having a size of 32 x 16, partitions 624 having a dimension of 16 x 32, and partitions 626 having a dimension of 16 x 16.

[000111] Similarly, a prediction unit of encoding unit 630 with the dimension 16 x 16 and the depth of 2 can be divided into partitions included in the encoding unit 630, that is, a partition with a dimension of 16 x16 included in encoding unit 630, partitions 632 having a dimension of 16x8, partitions 634 having a dimension of 8x16, and partitions 636 having a dimension of 8x8.

[000112] Similarly, a prediction unit of encoding unit 640 with a dimension of 8 x 8 and a depth of 3 can be divided into partitions included in the encoding unit 640, that is, a partition with a dimension of 8 x 8, included in encoding unit 640, partitions 642 having a dimension of 8 x 4, partitions 644 having a dimension of 4 x 8, and partitions 646 having a dimension of 4 x 4.

[000113] The 650 coding unit with the dimension 4 x 4 and the depth of 4 is the smallest coding unit and a lower depth coding unit. A prediction unit of encoding unit 650 is only assigned to a partition with a dimension of 4 x 4.

[000114] In order to determine the at least one encoded depth of the encoding units that make up the largest encoding unit 610, the hierarchical encoder 110 of the video encoding apparatus 100 performs encoding for the encoding units corresponding to each included depth in the larger coding unit 610.

[000115] The number of deeper encoding units according to depth including data in the same range and in the same dimension increases as the depth gets deeper. For example, four encoding units corresponding to a depth of 2 are needed to cover data that is included in one unit corresponding to a encoding depth of 1. Consequently, in order to compare the encoding results of the same data according to the depth, the encoding unit corresponding to the depth of 1 and four encoding units corresponding to the depth of 2 are each encoded.

[000116] In order to perform coding to a current depth between the depths, a minimum coding error can be selected for the current depth by performing the coding, for each prediction unit in the coding units corresponding to the current depth, along of the horizontal axis of the hierarchical structure 600. Alternatively, the minimum coding error can be searched for by comparing the minimum coding errors according to depth and can perform coding for each depth as the depth deepens along the vertical axis of hierarchical structure 600. A depth and a partition having the minimum encoding error in the largest encoding unit 610 can be selected as the encoded depth and a partition type of the largest encoding unit 610.

[000117] FIG. 7 is a diagram for describing a relationship between an encoding unit 710 and transforming units 720, in accordance with an embodiment of the present invention.

[000118] The video coding apparatus 100 or the video decoding apparatus 200 encodes or decodes a picture of each larger coding unit according to the coding units having dimensions equal to or smaller than the dimension of the larger coding unit . The dimensions of the transformation units for transformation during encoding can be selected based on data units that are no larger than a corresponding encoding unit.

[000119] For example, in the video encoding apparatus 100 or video decoding apparatus 200, if the dimension of the encoding unit 710 is 64 x 64, the transformation can be performed using the transforming units 720 with a dimension of 32 x 32.

[000120] In addition, the data of the encoding unit 710 with the dimension 64 x 64 can be encoded by performing the transformation in each of the transform units with the dimensions 32 x 32, 16 x 16, 8 x 8, and 4x4, which are smaller than 64x64, and then a transform unit having the minimum encoding error can be selected.

[000121] FIG. 8 is a diagram describing the encoding information of the encoding units corresponding to an encoded depth, in accordance with an embodiment of the present invention.

[000122] An output unit 130 of the video coding apparatus 100 can encode and transmit information 800 about a partition type, information 810 about a prediction mode, and information 820 about the size of a transformation unit for each encoding unit corresponding to an encoded depth, such as information about an encoding mode.

[000123] Information 800 indicates information about a shape of a partition obtained by dividing a prediction unit by a current encoding unit, wherein the partition is a prediction data unit encoding the current encoding unit. For example, a current CU_0 encoding unit with a dimension of 2N x 2N can be divided into any one of an 802 partition with a dimension of 2N x 2N, an 804 partition with a dimension of N x 2N, an 806 partition with a dimension. dimension of N x 2N, and an 808 partition with a dimension of N x N. Here, the 800 information about a partition type is set to indicate one of the 802 partition having a dimension of 2N x 2N, the 804 partition having a dimension of N x 2N, partition 806 having a dimension of N x 2N, and partition 808 having a dimension of N x N.

[000124] Information 810 indicates a prediction mode of each partition. For example, information 810 may indicate a prediction encoding mode performed on a partition indicated by information 800, i.e., an intra 812 mode, an inter 814 mode, or a skip 816 mode.

[000125] Information 820 indicates a transformation unit to be based on when the transformation is performed on a current encoding unit. For example, the transform unit can be a first intra 822 transform unit, a second intra 824 transform unit, a first inter 826 transform unit, or a second inter 828 transform unit.

[000126] The coding data and image data extracting unit 210 of the video decoding apparatus 200 can extract and use the information 800 about the coding units, the information 810 about a prediction mode, and the information 820 about a dimension of a transformation unit, for decoding, according to each deeper encoding unit.

[000127] FIG. 9 is a diagram of deeper coding units according to depth, in accordance with an embodiment of the present invention.

[000128] Split information can be used to indicate a change of a depth. The division information indicates whether a coding unit of a current depth is divided into coding units of a lesser depth.

[000129] A prediction unit 910 for prediction encoding a encoding unit 900 with a depth of 0 and a dimension of 2N_0 x 2N_0 may include partitions of a partition type 912 having a dimension of 2N_0 x 2N_0, a type of 914 partition having a dimension of 2N_0 x N_0, a partition type 916 having a dimension of N_0 x 2N_0, and a type 918 partition having a dimension of N_0 x N_0. FIG. 9 illustrates only partition types 912 to 918 that are obtained by symmetrically dividing prediction unit 910, but a partition type is not limited to them, and prediction unit 910 partitions can include asymmetric partitions, partitions with a shape predetermined, and the partitions with a geometric shape.

[000130] Prediction encoding is performed repeatedly on one partition with a dimension of 2N_0 x 2N_0, two partitions having a dimension of 2N_0 x N_0, two partitions having a dimension of N_0 x 2N_0, and four partitions having a dimension of N_0 x N_0, according to each partition type. Prediction coding in an intra mode and an inter mode can be performed on partitions having dimensions of 2N_0 x 2N_0, N_0 x 2N_0, 2N_0 x N_0, and N_0 x N_0. Prediction encoding in a skip mode is performed only on the partition with the dimension 2N_0 x 2N_0.

[000131] If a coding error is the smallest in one of the partition types 912 to 916 having dimensions 2N_0 x 2N_0, 2N_0 x N_0, and N_0 x 2N_0, the prediction unit 910 may not be split to a smaller depth .

[000132] If encoding error is the smallest in partition type 918 with dimension N_0 x N_0, depth is changed from 0 to 1 to divide partition type 918 in operation 920, and encoding is performed repeatedly in Encoding partition type units that have a depth of 2 and a dimension of N_0 x N_0 to look for a minimal encoding error.

[000133] A prediction unit 940 for prediction encoding of (partition type) encoding unit 930 with a depth of 1 and a dimension of 2N_1 x 2N_1 (= N_0 x N_0) may include partitions of a partition type 942 having a dimension of 2N_1 x 2N_1, a partition type 944 having a dimension of 2N_1 x N_1, a partition type 946 having a dimension of N_1 x 2N_1, and a partition type 948 having a dimension of N_1 x N_1.

[000134] If an encoding error is the smallest in partition type 948 with a dimension of N_1 x N_1, a depth is changed from 1 to 2 to divide partition type 948 in operation 950, and encoding is performed repeatedly on the coding units 960, having a depth of 2 and a dimension of N_2 x N_2 to search for a minimum coding error.

[000135] When a maximum depth is d, the division operation according to each depth can be performed even when a depth becomes d-1, and the division information can be encoded as even when a depth is one from 0 to d-2. In other words, when coding is performed until the depth is d-1 after a coding unit corresponding to a depth of d-2 is divided in operation 970, a prediction unit 990 for prediction coding of a unit of encoding 980 having a depth of d-1 and a dimension of 2N_(d-1)x2N_(d-1) may include partitions of a partition type 992 having a dimension of 2N_(d-1)x2N_(d-1) , a partition type 994 having a dimension of 2N_(d-1)xN_(d-1), a partition type 996 having a dimension of N_(d-1)x2N_(d-1), and a partition of the type 998 with a dimension of N_(d-1)xN_(d-1).

[000136] Prediction encoding can be performed repeatedly on one partition with a dimension of 2N_(d-1) x 2N_(d-1), two partitions having a dimension of 2N_(d-1) x N_(d-1 ), two partitions having a dimension of N_(d-1) x 2N_(d-1), four partitions having a dimension of N_(d-1) x N_(d-1), among the partition types 992 to 998 to search for a partition type having a minimal encoding error.

[000137] Even when partition type 998 with dimension N_(d-1) x N_(d-1) has the minimum encoding error, since a maximum depth is d, the encoding unit CU_(d -1) having a depth of d-1 is no longer divided to a smaller depth, and an encoded depth for the encoding units that make up the current largest encoding unit 900 is determined to be d-1 and a partition type of the current largest encoding unit 900 can be determined to be N_(d-1) x N_(d-1). Furthermore, since the maximum depth is d, the division information for the smallest encoding unit 952 is not defined.

[000138] Data unit 999 can be a ‘minimum unit’ for the current largest encoding unit. A minimum unit according to an embodiment of the present invention may be a rectangular data unit obtained by dividing the smallest encoding unit 980 by 4. When performing encoding repeatedly, the video encoding apparatus 100 may select a depth having the error of minimum coding, comparing the coding errors according to the depths of the coding unit 900 to determine a coded depth, and defining a corresponding partition type and a prediction mode as a coded depth coding mode.

[000139] As such, the minimum coding errors according to depths are compared at all depths from 1 to d, and a depth having the minimum coding error can be determined as a coded depth. The encoded depth, the prediction unit partition type, and the prediction mode can be encoded and transmitted as encoding mode information. Furthermore, since a coding unit is divided from a depth of 0 to a coded depth, only the coded depth division information is set to 0, and the depth division information excluding the coded depth is set to 1 .

[000140] The entropy decoder 220 of the video decoding apparatus 200 can extract and use the information about the encoded depth and the prediction unit of the encoding unit 900 to decode the encoding unit 912. The video decoding apparatus 200 can determine a depth, at which the division information is 0, as a coded depth using division information according to depth, and use the information about a corresponding depth coding mode for decoding.

[000141] FIGS. 10 to 12 are diagrams describing a relationship between encoding units 1010, prediction units 1060, and transformation units 1070 in accordance with an embodiment of the present invention.

[000142] Encoding units 1010 are encoding units having a tree structure corresponding to the encoded depths determined by the video encoding apparatus 100 in the larger encoding unit. Prediction units 1060 are partitions of prediction units from each of the coding units 1010, and the transformation units 1070 are transformation units from each of the coding units 1010.

[000143] When a depth of a larger coding unit is 0 in coding units 1010, the depths of coding units 1012 and 1054 are 1, the depths of coding units 1014, 1016, 1018, 1028, 1050, and 1052 are 2, the depths of the coding units 1020, 1022, 1024, 1026, 1030, 1032, and 1048 are 3, and the depths of the coding units 1040, 1042, 1044, and 1046 are 4.

[000144] In prediction units 1060, some coding units 1014, 1016, 1022, 1032, 1048, 1050, 1052, and 1054 are obtained by dividing the coding units. In other words, the partition types in encoding units 1014, 1022, 1050, and 1054 have a dimension of 2N x N, the partition types in encoding units 1016, 1048, and 1052 have a dimension of N x 2N, and a partition type of encoding unit 1032 has a dimension of N x N. Prediction units and partitions of encoding units 1010 are equal to or smaller than each encoding unit.

[000145] The inverse transformation or transformation is performed on the image data of the coding unit 1052 in the transformation units 1070 in a data unit that is smaller than the coding unit 1052. In addition, the coding units 1014, 1016 , 1022, 1032, 1048, 1050, 1052, and 1054 in transformation units 1070 are different from those in prediction units 1060 in terms of dimensions and shapes. In other words, the video coding apparatus 100 and the video decoding apparatus 200 can perform intraprediction, motion estimation, motion compensation, transform, and inverse transform individually on one data unit, in the same unit. of coding.

[000146] In this way, encoding is performed recursively in each of the encoding units having a hierarchical structure in each region of a larger encoding unit to determine an optimal encoding unit, and thus the encoding units having a recursive tree structure can be obtained. Encoding information can include division information about an encoding unit, information about a partition type, information about a prediction mode, and information about the size of a transformation unit.

[000147] Table 1 shows the encoding information that can be set by the video encoding apparatus 100 and the video encoding apparatus 200.

[000148] The entropy encoder 120 of the video decoding apparatus 100 can output the encoding information about the encoding units having a tree structure, and the entropy decoder 220 of the video decoding apparatus 200 can extract the information from encoding over encoding units having a tree structure of a received bit stream.

[000149] The division information indicates whether a current coding unit is divided into coding units of a smaller depth. If the division information of a current depth d is 0, a depth, at which a current encoding unit is no longer divided into a smaller depth, is an encoded depth, and so the information about a partition type, a mode of prediction, and a dimension of a transformation unit can be set to the coded depth. If the current coding unit is further divided according to the division information, the coding is performed independently into four division coding units of a smaller depth.

[000150] A prediction mode can be one of an intra mode, an inter mode, and a skip mode. Intra mode and inter mode can be set on all partition types, and skip mode is set only on a partition type with a dimension of 2N x 2N.

[000151] Partition type information can indicate symmetric partition types having dimensions of 2Nx2N, 2NxN, Nx2N, and NxN, which are obtained by symmetrically dividing a height or a width of a prediction unit, and the types of asymmetric partitions having dimensions of 2NxnU, 2NxnD, nLx2N, and nRx2N, which are obtained by asymmetrically dividing the height or width of the prediction unit. Asymmetric partition types having dimensions of 2NxnU and 2NxnD can be obtained, respectively, by dividing the height of the prediction unit into 1: nen:1 (where n is an integer greater than 1), and the types of Asymmetric partitions having the dimensions of nLx2N and nRx2N can be obtained, respectively, by dividing the width of the prediction unit into 1:nen:1.

[000152] The size of the transformation unit can be set to be of two types in intra mode and two types in inter mode. In other words, if the transform unit division information is 0, the transform unit dimension can be 2N x 2N, which is the current encoding unit dimension. If the transformation unit division information is 1, the transformation units can be obtained by dividing the current encoding unit. Also, if a current encoding unit partition type with a dimension of 2N x 2N is a symmetric partition type, a transformation unit dimension can be NxN, and if the current encoding unit partition type is an asymmetric partition type, the size of the transformation unit can be N/2 x N/2.

[000153] Encoding information about encoding units having a tree structure may include at least one of an encoding unit corresponding to an encoded depth, a prediction unit, and a minimum unit. The coding unit corresponding to the coded depth can include at least one of a prediction unit and a minimum unit containing the same coding information.

[000154] Thus, it is determined whether the adjacent data units are included in the same coding unit corresponding to the coded depth by comparing with the information coding of the adjacent data units. Furthermore, a coding unit corresponding to a corresponding coded depth is determined using the coding information of a data unit, and therefore a distribution of the coded depths in a larger coding unit can be determined.

[000155] Thus, if a current coding unit is predicted based on the coding information of the adjacent data units, the coding information of the data units in the deeper coding units adjacent to the current coding unit can be directly referenced and used.

[000156] Alternatively, if a current encoding unit is predicted based on the encoding information of the adjacent data units, the data units adjacent to the current encoding unit are searched using the encoded information of the data units, and the units searched adjacent encoding units can be referenced to predict the current encoding unit.

[000157] FIG. 13 is a diagram for describing a relationship between a coding unit, a prediction unit, and a transformation unit according to the coding mode information of Table 1.

[000158] A larger encoding unit 1300 includes the encoding units 1302, 1304, 1306, 1312, 1314, 1316, and 1318 of depth encoded. Here, since the coding unit 1318 is a coding unit of a coded depth, the division information can be set to 0. Information about a partition type of the coding unit 1318 having a dimension of 2N x 2N can be defined to be one of a 1322 partition type having a dimension of 2N x 2N, a 1324 partition type having a dimension of N x 2N, a 1326 partition type having a dimension of N x 2N, a 1328 partition type having a dimension of N x N, a partition type 1332 having a dimension of 2N x nU, a partition type 1334 having a dimension of 2N x nD, a partition type 1336 having a dimension of nL x 2N, and a partition of type 1338 having a dimension of nR x 2N.

[000159] When the partition type is defined to be symmetric, that is, partition type 1322, 1324, 1326, or 1328, a 1342 transformation unit having a dimension of 2N x 2N is defined if the split information (indicator of TU dimension) of a transformation unit is 0 and a 1344 transformation unit having a dimension of N x N is defined if a TU dimension indicator is 1.

[000160] When the partition type is set to be asymmetric, ie partition type 1332, 1334, 1336, or 1338, a 1352 transformation unit having a dimension of 2N x 2N is defined if a TU dimension indicator is 0, and a 1354 transform unit having a dimension of N/2 x N/2 is defined, if a TU dimension indicator is 1.

[000161] The TU dimension indicator is a type of transformation index; a dimension of a transformation unit corresponding to a transformation index can be modified according to a type of prediction unit or a partition type of an encoding unit.

[000162] When the partition type is set to be symmetric, ie partition type 1322, 1324, 1326, or 1328, the 1342 transformation unit having a dimension of 2N x 2N is defined if a TU dimension indicator of a transform unit is 0, and the transform unit 1344 having a dimension of N x N is defined if an indicator of dimension TU is 1.

[000163] When the partition type is set to be asymmetric, that is, partition type 1332 (2NxnU), 1334 (2NxnD), 1336 (nLx2N), or 1338 (nRx2N), the transformation unit 1352 with one dimension of 2N x 2N is defined if a TU dimension indicator is 0, and the 1354 transformer unit having a dimension of N/2 x N/2 is defined if a TU dimension indicator is 1.

[000164] With reference to FIG. 9, the TU size indicator described above is an indicator having a value of 0 or 1, but the TU size indicator is not limited to 1 bit, and a transform unit can be hierarchically divided while the TU size indicator increases by 0. Transformation unit division information (TU dimension indicator) can be used as an example of a transformation index.

[000165] In this case, when a TU dimension indicator, according to a modality, is used with a maximum dimension and a minimum dimension of a transformer unit, the dimension of the transformer unit actually used can be expressed. The video coding apparatus 100 can encode the size information of the larger transformation unit, the size information of the smaller transformation unit, and the division information of the larger transformation unit. The encoded largest transform unit dimension information, the smallest transform unit dimension information, and the largest transform unit split information can be entered into a sequence parameter set (SPS). The video decoding apparatus 200 can use the size information of the larger transformation unit, the size information of the smaller transformation unit, and the division information of the larger transformation unit for video decoding.

[000166] For example, (a) if a dimension of a current encoding unit is 64x64 and a larger transform unit is 32x32, (a-1) if a dimension of a transform unit is 32x32 if an indicator of dimension TU is 0; (a-2) if a dimension of a transformer unit is 16x16 if a TU dimension indicator is 1; and (a-3) if a transformation unit dimension is 8x8 if a TU dimension indicator is 2.

[000167] Alternatively, (b) if a dimension of a current coding unit is 32x32 and a smaller transform unit is 32x32, (b-1) if a dimension of a transform unit is 32x32 if a TU dimension indicator is 0, and since the dimension of a transformer unit cannot be smaller than 32 x 32, no other TU dimension indicator can be defined.

[000168] Alternatively, (c) if a dimension of a current encoding unit is 64x64 and a maximum TU dimension indicator is 1, a TU dimension indicator can be 0 or 1 and no other TU size indicator can be defined.

[000169] Consequently, when defining a maximum TU dimension indicator 'MaxTransformSizeIndex', a minimum TU dimension indicator as 'MinTransformSize', and a transformation unit in the case where a TU dimension indicator is 0, ie, RootTu root transform unit as 'RootTuSize ', a dimension of a smaller transform unit 'CurrMinTuSize', which is available in a current encoding unit, can be defined by equation (1) below. CurrMinTuSize = max (MinTransformSize, RootTuSize/(2AMaxTransformSizeIndex)) (1)

[000170] Compared to the smaller transformation unit dimension 'CurrMinTuSize' which is available in the current coding unit, the root transformation unit dimension 'RootTuSize', which is a dimension of a transformation unit when using an indicator of dimension TU is 0, it may indicate a larger transformation unit that can be selected with respect to a system. That is, according to Equation (1), 'RootTuSize/(2AMaxTransformSizeIndex)' is a dimension of a transformation unit that is obtained by dividing 'RootTuSize', which is a dimension of a transformation unit when the information of transformation unit division is 0, by the number of division times corresponding to the division information of the larger transformation unit, and 'MinTransformSize' is a dimension of a smaller transformation unit, so a smaller value of these can be 'CurrMinTuSize' ', which is the dimension of the smallest transformation unit that is available in the current encoding unit.

[000171] The size of the transformation unit in the root 'RootTuSize' according to an embodiment of the present invention may vary according to a prediction mode.

[000172] For example, if a current prediction mode is an inter mode, RootTuSize can be determined according to equation (2) below. In equation (2), 'MaxTransformSize' refers to a larger transformation unit dimension, and 'PUSize' refers to a current prediction unit dimension. RootTuSize = min(MaxTransformSize, PUSize) (2)

[000173] In other words, if a current prediction mode is an inter mode, the dimension dimension of the root transformation unit 'RootTuSize', which is a transformation unit if a TU dimension indicator is 0, can be set as a smaller value between the dimension of the largest transformation unit and the dimension of the current prediction unit.

[000174] If a prediction mode of a current partition unit is an intra mode, 'RootTuSize' can be determined according to Equation (3) below. 'PartitionSize' refers to a dimension of the current partition drive. RootTuSize = min (MaxTransformSize, PartitionSize) (3)

[000175] In other words, if a current prediction mode is an intra mode, the root transformation unit dimension 'RootTuSize' can be set to a smaller value between the larger transformation unit dimension and the root transformation unit dimension. current partition.

[000176] However, it should be noted that the dimension of the root transform unit dimension 'RootTuSize', which is the dimension of the current largest transform unit according to an embodiment of the present invention and varies according to a mode prediction of a partition unit, is an example, and factors for determining the size of the current largest transformation unit are not limited to these.

[000177] An entropy encoding operation of a syntax element, which is performed by the entropy encoder 120 of the video encoding apparatus 100 of FIG. 1, and an entropy decoding operation of a syntax element that is performed by the entropy decoder 220 of the video decoding apparatus 200 of FIG. 2, will now be described in detail.

[000178] As described above, the video encoding apparatus 100 and the video decoding apparatus 200 perform encoding and decoding by dividing a larger encoding unit into encoding units that are equal to or smaller than the encoding unit. higher encoding. A prediction unit and a transformation unit used in prediction and transformation can be determined on a cost basis independently of the other data units. Since an optimal encoding unit can be determined by recursively encoding each encoding unit having a hierarchical structure included in the larger encoding unit, data units having a tree structure can be configured. In other words, for each larger encoding unit, an encoding unit having a tree structure and a prediction unit and a transformation unit each having a tree structure can be configured. For decoding, hierarchical information, which is information that indicates the structure information of data units having a hierarchical structure, and non-hierarchical information for decoding, in addition to hierarchical information, need to be transmitted.

[000179] Information related to a hierarchical structure is information necessary to determine a coding unit having a tree structure, and a prediction unit having a tree structure, and a transformation unit having a tree structure, as described above with reference to FIGS. 10 to 12, and includes size information of a larger coding unit, coded depth, partition information of a prediction unit, a division indicator indicating whether a coding unit is divided, size information of a transformation unit , and a split transform indicator (split_transform_flag) indicating whether a transform unit is split into smaller transform units for a transform operation. Examples of coding information other than hierarchical structure include inter/intraprediction prediction mode information applied to each prediction unit, motion vector information, prediction direction information, color component information applied to each data unit in the case where a plurality of color components is used and the transformation coefficient level information. Hereinafter, hierarchical information and extra-hierarchical information can be referred to as a syntax element that must be entropy encoded or entropy decoded.

[000180] In particular, according to embodiments of the present invention, a method to select a context model when a syntax element related to a transformation unit among the syntax elements is provided. An operation of entropy encoding and decoding of syntax elements related to a transformation unit will now be described in detail.

[000181] FIG. 14 is a block diagram of an entropy encoding apparatus 1400 in accordance with an embodiment of the present invention. Entropy encoding apparatus 1400 corresponds to entropy encoder 120 of video encoding apparatus 100 of FIG. 1.

[000182] Referring to FIG. 14, the entropy encoding apparatus 1400 includes a binarizer 1410, a context modeler 1420, and a binary arithmetic encoder 1430. In addition, the binary arithmetic encoder 1430 includes a regular encoding mechanism 1432 and a bypass encoding mechanism 1434.

[000183] When the syntax elements input to the entropy coding apparatus 1400 are not binary values, the binarizer 1410 binarizes the syntax elements to produce a string of bins consisting of binary values of 0 and 1. A bin denotes each bit of a stream consisting of 0 and 1, and is encoded by context adaptive binary arithmetic encoding (CABAC). If a syntax element is given having the same probability between 0 and 1, the syntax element is output to bypass encoding mechanism 1434, which does not use a probability, to be encoded.

[000184] Binarizer 1410 can use various binarization methods according to the type of a syntax element. Examples of binarization methods can include a unary method, a truncated unary method, a truncated rice code method, a Golomb code method, and a fixed-length code method.

[000185] A cbf Transform Unit Significant Coefficient Indicator that indicates whether a non-zero transform coefficient (hereinafter also referred to as a "significant coefficient") exists in a transform unit is binarized using the fixed code method. That is, if a non-zero transformation coefficient exists in the transformation unit, the transformation unit significant coefficient indicator cbf is defined as having a value of 1. Otherwise, if a non-zero transformation coefficient does not exist in the transform unit, the transform unit significant coefficient indicator cbf is set to have a value equal to 0. If an image includes a plurality of color components, the transform unit significant coefficient indicator cbf can be set with respect to a transformation unit of each color component. For example, if an image includes luminance (Y) and chrominance (Cb, Cr) components, a transform unit significant coefficient indicator cbf_luma of a luminance component transform unit, and a unit significant coefficient indicator. cbf_cb or cbf_cr transformation of the chrominance component transformation unit can be defined.

[000186] The context modeler 1420 provides a context model for encoding a bit string corresponding to a syntax element, for the regular encoding mechanism 1432. In more detail, the context modeler 1420 produces a probability of one binary value for encoding each binary value of a bit string of a current syntax element, to binary arithmetic encoder 1430.

[000187] A context model is a probability model of a bin, and includes information on which of 0 and 1 corresponds to a most likely symbol (MPS) and a least likely symbol (LPS), and the probability information of at least one. minus one among MPS and the LPS.

[000188] The context modeler 1420 can select a context model for the entropy coding of the cbf transformation unit significant coefficient indicator, based on a transformation unit depth. If the dimension of the transform unit is equal to the dimension of one encoding unit, that is, if the transform depth of the transform unit is 0, the context modeler 1420 can determine a first predefined context model as a context model for the entropy coding of the cbf transformation unit significant coefficient indicator. Otherwise, if the size of the transformation unit is smaller than the size of the encoding unit, that is, if the transformation unit depth is not 0, the context modeler 1420 can determine a second predefined context model as a context model for entropy coding of the cbf transformation unit significant coefficient indicator. Here, the first and second context models are based on different probability distribution models. That is, the first and second context models are different context models.

[000189] As described above, when the cbf transformation unit significant coefficient indicator is encoded by entropy, the context modeler 1420 uses different context models in a case where the transformation unit dimension equals the unit dimension coding, and a case where the size of the processing unit is not equal to the size of the coding unit. If an index indicating one of a plurality of predefined context models for entropy encoding the cbf transformation unit significant coefficient indicator is referred to as a context index ctxIdx, the context index ctxIdx may have a value obtained by the sum of a ctxInc context boost parameter to determine a context template, and a ctxIdxOffset predefined context index offset. That is, ctxIdx = ctxInc + ctxIdxOffset. The context modeler 1420 can distinguish a case where the transformation unit transformation depth is 0 from a case where the transformation unit transformation depth is not 0, it can change the context boost parameter ctxInc to determine a context model, from the transformation depth of the transformation unit, and thus can change the context index ctxIdx to determine a context model for the entropy coding of the indicator of significant coefficient of transformation unit cbf.

[000190] In more detail, if the transformation depth is referred to as trafodepth, the context modeler 1420 can determine the context augmentation parameter ctxInc based on the following algorithm. ctxInc = (trafodepth == 0)? 1: 0

[000191] This algorithm can be implemented by the following pseudocode. { If (trafodepth==0) ctxInc=1; else ctxInc=0; }

[000192] The cbf transformation unit significant coefficient indicator can be set separately according to the luminance and chrominance components. As described above, a context model for entropy coding of the luminance component transform unit cbf_luma transform unit significant coefficient indicator can be determined using the context boost parameter ctxInc which changes depending on whether the transformation depth of the transform unit is 0. A context model for entropy coding of the cbf_cb transform unit or cbf_cr transform unit significant coefficient indicator of the chrominance component transform unit can be determined using a trafodepth transform depth value as the increment parameter ctxInc.

[000193] Regular encoding engine 1432 performs binary arithmetic encoding on a bit stream corresponding to a syntax element, based on information about the MPS and LPS and the probability information of at least one of the MPS and the LPS, which are included in the scope of the model provided from the 1420 context modeler.

[000194] FIG. 15 is a flowchart of an entropy encoding and decoding operation of a syntax element related to a transformation unit, in accordance with an embodiment of the present invention.

[000195] With reference to FIG. 15, in operation 1510, a cbf transform unit significant coefficient indicator that indicates whether a non-zero transform coefficient exists among the transform coefficients included in a current transform unit is initially encoded and decoded by entropy. As described above, a context model for entropy coding of the cbf transformation unit significant coefficient indicator can be determined based on a transformation unit transformation depth and binary arithmetic coding on the cbf unit significant coefficient indicator. cbf transformation can be performed based on given context model.

[000196] If the cbf transformation unit significant coefficient indicator is 0, since only transformation coefficients of 0 exist in the current transformation unit, only a value 0 is entropy encoded or decoded as the significant coefficient indicator of cbf transform unit, and the transform coefficient level information is not encoded or decoded by entropy.

[000197] In operation 1520, if a significant coefficient exists in the current transform unit, a SigMap significance map indicating a location of a significant coefficient is encoded or decoded by entropy.

[000198] A SigMap significance map can be formed from a significant bit and predetermined information indicating the location of a last significance coefficient. A significant bit indicates whether a transformation coefficient according to each digitization index is a significant coefficient or 0, and can be expressed by significant_coeff_flag[i]. As will be described below, a significance map is defined in subset units having a predetermined dimension which is obtained by dividing the transformation unit. Consequently, significant_coeff_flag[i] indicates whether a transformation coefficient of an ith digitization index among the transformation coefficients included in a subset included in the transformation unit is 0.

[000199] According to conventional H.264, an indicator (End-Of-Block) that indicates whether each significant coefficient is the last significant coefficient is separately encoded or entropy decoded. However, according to an embodiment of the present invention, the location information of the last significant coefficient itself is encoded or decoded by entropy. For example, if a location of the last significant coefficient is (x, y), where x and y are integers, and last_significant_coeff_x and last_significant_coeff_y which are the syntax elements indicating coordinate values of (x, y) can be encoded or decoded by entropy.

[000200] In operation 1530, the transform coefficient level information indicating a dimension of a transform coefficient is encoded or decoded by entropy. According to conventional H.264/AVC, the transformation coefficient level information is expressed by coeff_abs_level_minus1 which is a syntax element. According to the embodiments of the present invention, as the transformation coefficient level information, coeff_abs_level_greater1_flag which is a syntax element about whether an absolute value of a transformation coefficient is greater than 1, coeff_abs_level_greater2_flag which is a syntax element about if an absolute value of a transformation coefficient is greater than 2, and coeff_abs_level_remaining which indicates the dimension information of the remaining transformation coefficient are encoded.

[000201] The coeff_abs_level_remaining syntax element indicating the remaining transform coefficient dimension information has a difference in range between a transform coefficient dimension (absCoeff) and a BaseLevel base level value, which is determined using coeff_abs_level_greater1_flag and coeff_abs_level_greater2_flag . The BaseLevel base level value is determined according to the equation: baseLevel=1+coeff_abs_level_greather1_flag + coeff_abs_level_greather2_flag and coeff_abs_level_remaining is determined according to the equation: coeff_abs_level_remaining=absCoeff-baseLevel. While coeff_abs_level_greater1_flag and coeff_abs_level_greater2_flag have a value of 0 or 1, the BaseLevel base level value can have a value of 1 to 3. Thus, coeff_abs_level_remaining can be varied from (absCoeff-1) to (absCoeff-3). As described above, (absCoeff-BaseLevel), which is a difference between the size of an original absCoeff transformation coefficient and the BaseLevel base level value, is transmitted as dimension information of a transformation coefficient in order to reduce the dimension of the data transmitted.

[000202] An operation of determining a context model for entropy encoding a transform unit significant coefficient indicator, according to an embodiment of the present invention, will now be described.

[000203] FIG. 16 is a diagram illustrating an encoding unit and transforming units 1611 to 1617 included in the encoding unit, according to an embodiment of the present invention. In FIG. 16, a data unit indicated by a dashed line denotes the coding unit, and data units indicated by solid lines indicate the transform units 1611 to 1617.

[000204] As described above, the video encoding apparatus 100 and the video decoding apparatus 200 perform encoding and decoding by dividing a larger encoding unit into encoding units having a dimension equal to or smaller than the dimension. of the larger encoding unit. A prediction unit and a transformation unit used in a prediction operation and a transformation operation can be determined on a cost basis independently of other data units. If the dimension of an encoding unit is larger than the dimension of a larger transforming unit usable by the video encoding apparatus 100 and the video encoding apparatus 200, the encoding unit can be divided into transforming units with a dimension equal to or smaller than the size of the largest transformation unit, and a transformation operation can be performed based on the division transformation units. For example, if the dimension of an encoding unit is 64x64 and the dimension of a usable largest transform unit is 32 x 32, in order to transform (or inversely transform) the encoding unit, the encoding unit is split in transformer units having a dimension equal to or smaller than 32 x 32.

[000205] A trafodepth indicating the number of times the coding unit is divided into horizontal and vertical directions in the transform units can be determined. For example, if the dimension of a current coding unit is 2Nx2N and the dimension of the transformation unit is 2Nx2N, the transformation depth can be set to 0. If the transformation unit dimension is NxN, the transformation depth can be set to 1. Otherwise, if the dimension of the transformation unit is N/2 x N/2, the transformation depth can be set to 2.

[000206] Referring to FIG. 16, transform units 1611, 1616, and 1617 are level 1 transform units obtained by dividing a root coding unit once, and have a transform depth of 1. Transform units 1612, 1614, 1614, and 1615 are the tier 2 transform units obtained by dividing a tier 1 transform unit into four pieces, and have a transform depth of 2.

[000207] FIG. 17 is a diagram illustrating a context enhancement parameter ctxInc used to determine a context model of a cbf transform unit significant coefficient indicator of each of transform units 1611 to 1617 of FIG. 16, based on a transformation depth. In the tree structure of FIG. 17, leaf nodes 1711 to 1717, respectively, correspond to transformation units 1611 to 1617 of FIG. 16, and the values of 0 and 1 marked on leaf nodes 1711 to 1717 indicate the cbf transform unit significant coefficient indicator of transform units 1611 to 1617. Further, in FIG. 17, leaf nodes with the same transformation depth are illustrated in the order of the transformation units located in the upper left, upper right, lower left, and lower right. For example, leaf nodes 1712, 1713, 1714, and 1715 of FIG. 17 correspond, respectively, to the transforming units 1612, 1613, 1614, and 1615 of FIG. 16. Furthermore, with reference to FIGS. 16 and 17, it is assumed that only the cbf transformation unit significant coefficient indicators of the transformation units 1612 and 1614 are 1, and that the cbf transformation unit significant coefficient indicators of the other transformation units are 0.

[000208] With reference to FIG. 17, since all of the transformation units 1611 to 1617 of FIG. 16 are obtained by dividing a root coding unit and therefore have non-zero transformation depths, the context boost parameter ctxInc used to determine a context model of the cbf transformation unit significant coefficient indicator of each one of the transform units 1611 to 1617 is defined to have a value of 0.

[000209] FIG. 18 is a diagram illustrating an encoding unit 1811 and a transforming unit 1812 included in encoding unit 1811, in accordance with another embodiment of the present invention. In FIG. 18, a data unit indicated by a dashed line indicates the encoding unit 1811, and a data unit indicated by a solid line indicates the transform unit 1812.

[000210] Referring to FIG. 18, if the dimension of the encoding unit 1811 is equal to the dimension of the transforming unit 1812 used to transform the encoding unit 1811, a trafodepth of the transforming unit 1812 has a value equal to 0. If the unit transform 1812 has a transform depth of 0, a context boost parameter ctxInc used to determine a context model of a transform unit cbf significant coefficient indicator of transform unit 1812 is set to have a value of 1.

[000211] The context modeler 1420 of FIG. 14 can compare the dimension of a coding unit to the dimension of a transformation unit based on a transformation unit's transformation depth, can distinguish a case where the transformation unit's transformation depth is 0 from a in which case the transformation unit transformation depth is not 0, and thus can change the context augmentation parameter ctxInc used to determine a context model for entropy coding the transformation unit significant coefficient indicator cbf . By changing the ctxInc context boost parameter used to determine a context model, the context model for entropy coding of the cbf transformation unit significant coefficient indicator can be changed in a case where the transformation unit depth of transformation is 0 and a case where the transformation depth of the transformation unit is not 0.

[000212] FIG. 19 is a diagram illustrating split_transform_flag split transform indicators used to determine the structure of the transform units included in the encoding unit of FIG. 16, in accordance with an embodiment of the present invention.

[000213] The video coding apparatus 100 can signal information about the structure of the transforming units used to transform each coding unit, to the video decoding apparatus 200. The information about the structure of the transforming units can be signaled by means the split transform indicator split_transform_flag which indicates whether each encoding unit is divided into horizontal and vertical directions into four transform units.

[000214] Referring to FIGS. 16 and 19, since a root encoding unit is divided into four pieces, a split_transform_flag 1910 split transform indicator of the root encoding unit is set to 1. If the dimension of the root encoding unit is greater than than the dimension of a larger usable transform unit, the split_transform_flag 1910 split transform indicator of the root encoding unit can always be set to 1 and cannot be flagged. This is because, if the size of a coding unit is larger than the size of a larger usable transformation unit, the coding unit need not be divided into deeper coding units that have a size equal to or less than the dimension of at least one larger processing unit.

[000215] With respect to each of the four divisions of the transformation units of the root coding unit that had a transformation depth of 1, a division transformation indicator that indicates whether the division of each of the four transformation units into four transformation units having a transformation depth of 2 is defined. In FIG. 19, split transform indicators of the transforming units having the same transforming depth are illustrated in the order of the transforming units located in the upper left, upper right, lower left, and lower right. A reference numeral 1911 denotes a division transform indicator of the transform unit 1611 of FIG. 16. Since the transforming unit 1611 is not divided into transforming units having a smaller depth, the transforming division indicator 1911 of the transforming unit 1611 has a value equal to 0. Likewise, since the units transforms 1616 and 1617 of FIG. 16 are not divided into transformation units having a smaller depth, the division transformation indicators 1913 and 1914 of the transformation units 1616 and 1617 have a value of 0. Since the transformation unit in the upper right having a depth of transformation 1 in FIG. 16 is divided into transforming units 1612, 1613, 1614, and 1615 having a transforming depth of 2, a division transform indicator 1912 of the top right transforming unit has a transforming depth of 1. Since the transformation units 1612, 1613, 1614, and 1615 with a transformation depth of 2 are not divided into transformation units with a smaller depth, division transformation indicators 1915, 1916, 1917 and 1918 of the transformation units 1612, 1613 , 1614 and 1615 having a transformation depth of 2 have a value of 0.

[000216] As described above, a context model for entropy coding of a cbf transformation unit significant coefficient indicator can be determined based on a transformation depth of a transformation unit, and binary arithmetic coding can be performed in the transform unit significant coefficient indicator based on the selected model context. If the cbf transformation unit significant coefficient indicator is 0, since only 0 transformation coefficients exist in a current transformation unit, only a value of 0 is entropy encoded or decoded as the unit significant coefficient indicator. cbf transformation, and the transformation coefficient level information is not entropy encoded or decoded.

[000217] An entropy encoding operation of a syntax element related to transformation coefficients included in a transformation unit of which a cbf transformation unit significant coefficient indicator has a value of 1, that is, a transformation unit having a non-zero transformation coefficient will now be described.

[000218] FIG. 20 illustrates a transform unit 2000 that is entropy encoded in accordance with an embodiment of the present invention. Although the transformer unit 2000 having a dimension of 16 x 16 is illustrated in FIG. 20, the dimension of the transformer unit 2000 is not limited to the illustrated dimension of 16x16, but can also be of various dimensions ranging from 4x4 to 32x32.

[000219] With reference to FIG. 20, for the entropy encoding and decoding of the transformation coefficient included in the transformation unit 2000, the transformation unit 2000 can be divided into smaller transformation units. An entropy encoding operation of a syntax element related to a 2010 4x4 Transformer Unit included in Transformer Unit 2000 will now be described. This entropy encoding operation can also be applied to a transformer unit of different dimensions.

[000220] The transformation coefficients included in the 2010 4x4 transformation unit each have a transformation coefficient (absCoeff) as illustrated in FIG. 20. The transform coefficients included in the 2010 4x4 Transform Unit can be serialized according to a predetermined digitizing order as illustrated in FIG. 20 and processed sequentially. However, the scan order is not limited as illustrated, but can also be modified.

[000221] Examples of syntax elements related to the transformation coefficients included in the 2010 4x4 transformation unit are significant_coeff_flag which is a syntax element indicating whether each transformation coefficient included in a transformation unit is a significant coefficient that has a value which is not 0, which is a coeff_abs_level_greater1_flag which is a syntax element indicating whether an absolute value of the transformation coefficient is greater than 1, coeff_abs_level_greater2_flag which is a syntax element indicating whether the absolute value is greater than 2, and coeff_abs_level_remaining which is a syntax element indicating the dimension information of the remaining transformation coefficients.

[000222] FIG. 21 illustrates a SigMap 2100 significance map corresponding to the 2010 transformation unit of FIG. 20.

[000223] Referring to FIGS. 20 and 21, the SigMap 2100 significance map having a value of 1 for each of the significant coefficients having a value other than 0, and among the transformation coefficients included in the 2010 4x4 transform unit of FIG. 20, is defined. The SigMap 2100 significance map is entropy encoded or decoded using a predefined context model.

[000224] FIG. 22 illustrates coeff_abs_level_greater1_flag 2200 corresponding to the 2010 4x4 transform unit of FIG. 20.

[000225] Referring to FIGS. 20 to 22, the coeff_abs_level_greater1_flag 2200 which is an indicator indicating whether a corresponding significance transformation coefficient has a value greater than 1, over significant coefficients for which the SigMap 2100 significance map has a value of 1, is defined. When coeff_abs_level_greater1_flag 2200 is 1, it indicates that a corresponding transformation coefficient is a transformation coefficient having a value greater than 1, and when coeff_abs_level_greater1_flag 2200 is 0, it indicates that a corresponding transformation coefficient is a transformation coefficient with a value of 1. In FIG. 22, when coeff_abs_level_greater1_flag 2210 is at a location of a transform coefficient having a value of 1, the coeff_abs_level_greater1_flag 2210 has a value of 0.

[000226] FIG. 23 illustrates coeff_abs_level_greater2_flag 2300 corresponding to the 2010 4x4 transform unit of FIG. 20.

[000227] Referring to FIGS. 20 to 23, coeff_abs_level_greater2_flag 2300 indicating whether a corresponding transform coefficient has a value greater than 2, with respect to the transform coefficients for which coeff_abs_level_greater1_flag 2200 is set to 1, is defined. When coeff_abs_level_greater2_flag 2300 is 1, it indicates that a corresponding transformation coefficient is a transformation coefficient having a value greater than 2, and when coeff_abs_level_greater2_flag 2300 is 0, it indicates that a corresponding transformation coefficient is a transformation coefficient with a value of 2. In FIG. 23, when coeff_abs_level_greater2_flag 2310 is at a transform coefficient location with a value of 2, the coeff_abs_level_greater2_flag 2310 has a value of 0.

[000228] FIG. 24 illustrates the coeff_abs_level_remaining 2400 corresponding to the 2010 4x4 transform unit of FIG. 20.

[000229] Referring to FIGS. 20 to 24, the coeff_abs_level_remaining 2400, which is a syntax element indicating dimension information of the remaining transformation coefficients, can be obtained by calculating (absCoeff-BaseLevel) of each transformation coefficient.

[000230] The coeff_abs_level_remaining 2400 which is the syntax element indicating the dimension information of the remaining transform coefficients has a difference in one range between the dimension of the transform coefficient (absCoeff) and a BaseLevel base level value determined using coeff_abs_level_greater1_flag e coeff_abs_level_greater2_flag. The BaseLevel base level value is determined according to the equation: baseLevel=1+coeff_abs_level_greather1_flag + coeff_abs_level_greather2_flag and coeff_abs_level_remaining is determined according to the equation: coeff_abs_level_remaining=absCoeff-baseLevel.

[000231] The coeff_abs_level_remaining 2400 can be read and encoded by entropy according to the illustrated digitization order.

[000232] FIG. 25 is a flowchart of a method of entropy coding a video, in accordance with an embodiment of the present invention.

[000233] Referring to FIGS. 14 and 25, in operation 2510, context modeler 1420 obtains data from an encoding unit transformed on the basis of a transforming unit. In operation 2520, context modeler 1420 determines a context model to arithmetically encode a transformation unit significant coefficient indicator that indicates whether a non-zero transformation coefficient exists in the transformation unit, based on a transformation depth of the processing unit.

[000234] The context modeler 1420 can determine different context models, in a case where the size of the transformation unit is equal to the size of the encoding unit, that is, when the transformation depth of the transformation unit is 0, and a case where the size of the transformation unit is smaller than the size of the encoding unit, that is, when the transformation unit depth is not 0. In more detail, the context modeler 1420 can change a ctxInc context boost parameter to determine a context model, based on the transformation unit depth of transformation, can distinguish a case where the transformation unit transformation depth is 0 from a transformation case where the transformation unit depth is transform unit is not 0, and thus can change a context index ctxIdx to determine a context model for entropy encoding a coefficient indicator and significant transformation unit.

[000235] The significant coefficient indicator of the transformer unit can be set separately according to the luminance and chrominance components. A context model for entropy encoding a significant coefficient indicator of the luminance component transform unit cbf_luma transform unit can be determined using the context increase parameter ctxInc changed depending on whether the transform unit transformation depth is 0. A context model for entropy encoding a chrominance component transform unit cbf_cb or cbf_cr transform unit significant coefficient indicator can be determined using a trafodepth value as the increment parameter ctxInc context.

[000236] In operation 2530, the regular encoding engine 1432 arithmetically encodes the transform unit significant coefficient indicator based on the given context model.

[000237] FIG. 26 is a block diagram of an entropy decoding apparatus 2600 in accordance with an embodiment of the present invention. The entropy decoding apparatus 2600 corresponds to the entropy decoder 220 of the video decoding apparatus 200 of FIG. 2. The entropy decoding apparatus 2600 performs an inverse operation to the entropy encoding operation performed by the entropy encoding apparatus 1400 described above.

[000238] With reference to FIG. 26, entropy decoding apparatus 2600 includes a context modeler 2610, a regular decoding mechanism 2620, a bypass decoding mechanism 2630, and a debinarizer 2640.

[000239] A syntax element encoded using bypass encoding is produced for bypass decoder 2630 to be arithmetically decoded, and a syntax element encoded using normal encoding is arithmetically decoded by regular decoder 2620. Regular decoder 2620 decodes arithmetically a binary value of a current syntax element based on a context model provided using the context modeler 2610 to thereby produce a string of bits.

[000240] As the above-described context modeler 1420 of FIG. 14, the context modeler 2610 can select a context model to entropy decode a transform unit significant coefficient indicator cbf, based on a transform depth of one transform unit. That is, the context modeler 2610 can determine different context models, in a case where the size of the transformation unit is equal to the size of a coding unit, that is, when the transformation depth of the transformation unit is 0 , and a case where the size of the transformation unit is smaller than the size of the encoding unit, that is, when the transformation depth of the transformation unit is not 0. In more detail, the context modeler 2610 can change a ctxInc context boost parameter to determine a context model, based on the transformation unit depth of transformation, can distinguish a case where the transformation unit transformation depth is 0 from a case where the transformation unit depth is The transformation unit's transformation value is not 0, and thus can change a context index ctxIdx for determining a context model for entropy decoding of the information. cbf transformation unit significant coefficient indicator.

[000241] If the structure of the transformation units included in a coding unit is determined based on a split transform indicator split_transform_flag a if a coding unit obtained from a bit stream is divided into transformation units, the Transformation depth of the transformation unit can be determined based on the number of times the encoding unit is divided in order to reach the transformation unit.

[000242] The cbf transformation unit significant coefficient indicator can be set separately according to the luminance and chrominance components. A context model to entropy decode a significant coefficient indicator of the luminance component transform unit cbf_luma transform unit can be determined using the context boost parameter ctxInc which changes depending on whether the transform unit transformation depth is 0. A context model for entropy decoding of a chrominance component transform unit cbf_cb or cbf_cr transform unit significant coefficient indicator can be determined using a trafodepth value as the increment parameter ctxInc context.

[000243] The debinarizer 2640 reconstructs the bit strings that are arithmetically decoded by the regular decoding mechanism 2620 or bypass decoding mechanism 2630, into the syntax elements again.

[000244] The entropy decoding apparatus 2600 arithmetically decodes the syntax elements related to the transformation units, such as coeff_abs_level_remaing, SigMap, coeff_abs_level_greater1_flag, and coeff_abs_level_greater2_flag in addition to the same unit-significant cbf and unit indicator . When the syntax elements related to a transformation unit are reconstructed, the data included in the transformation units can be decoded using inverse quantization, inverse transformation and predictive decoding, based on the reconstructed syntax elements

[000245] FIG. 27 is a flowchart of a method of entropy decoding a video, in accordance with an embodiment of the present invention.

[000246] With reference to FIG. 27, in operation 2710, a transformation unit included in an encoding unit and used to inversely transform the encoding unit is determined. As described above, the structure of the transform units included in an encoding unit can be determined based on a split_transform_flag split transform indicator that indicates whether the encoding unit obtained from a bit stream is divided into transform units . In addition, a transformation unit depth can be determined based on the number of times the encoding unit is divided to reach the transformation unit.

[000247] In operation 2720, the context modeler 2610 obtains a transform unit significant coefficient indicator that indicates whether a non-zero transform coefficient exists in the transform unit, from the bit stream.

[000248] In operation 2730, context modeler 2610 determines a context model to arithmetically decode the transform unit's significant coefficient indicator, based on the transform unit's transform depth. As described above, the context modeler 2610 can determine different context models in a case where the size of the transform unit is equal to the dimension of the encoding unit, that is, when the transform depth of the transform unit is 0, and a case where the size of the transformation unit is smaller than the size of the encoding unit, that is, when the transformation unit depth is not 0. In more detail, the context modeler 2610 can change a ctxInc context boost parameter for determining a context model, based on the transformation unit's transformation depth, can distinguish a case where the transformation unit's transformation depth is 0 from a case where the transformation depth of the transformation unit is not 0 and thus can change a context index ctxIdx to determine a context model for entropy decoding. that of the significant coefficient indicator of the processing unit.

[000249] In operation 2740, the regular decoding engine 2620 arithmetically decodes the transform unit significant coefficient indicator based on the context model provided from the context modeler 2610.

[000250] Prior embodiments of the present invention may also be incorporated as computer readable code in a computer readable recording medium. Computer readable recording medium is any data storage device that can store data that can be later read by a computer system. Examples of computer-readable recording media include read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disks, and optical data storage devices. The computer readable recording medium can also be distributed by computer systems coupled to the network so that the computer readable code is stored and executed in a distributed manner.

[000251] Although the present invention has been presented and described particularly with reference to exemplary embodiments thereof, it will be understood by a person skilled in the art that various changes in form and details can be made therein without departing from the spirit and scope of the present invention as defined by the following claims.

Claims (3)

1. METHOD OF DECODING A VIDEO, characterized in that it comprises: obtaining (2720), from a bit stream, a division transformation indicator for a current depth and, when the division transformation indicator indicates no division for the current depth, determine that a transformation depth is equal to the current depth; obtaining at least one transform unit from an encoding unit based on the division transform indicator; determine (2730) a context augmentation parameter to determine a context index based only on whether the transformation depth is equal to zero or not, and obtain a context model to arithmetically decode a significant coefficient indicator of the transformation unit of a luma component indicating a non-zero transform coefficient exists in the transform unit of the luma component, using the context index obtained by adding the context boost parameter and a context shift; arithmetically decode (2740) the significant coefficient indicator of the transformation unit based on the given context model; and determining whether at least one non-zero transformation coefficient exists in the transformation unit's depth of transformation, based on the transformation unit's significant coefficient indicator, where the transformation depth indicates the number of times the coding unit is divided to determine the transform unit included in the encoding unit, wherein determining the context augmentation parameter further comprises: when the transform depth is equal to zero, determining the context augmentation parameter as one; and when the transformation depth is not equal to zero, set the context boost parameter to zero, when the division transformation indicator for the current transformation unit of the current transformation depth k indicates a division, the transformation unit of the Current transformation depth k is divided into four transformation units with a lower transformation depth (k + 1). 2. A VIDEO ENCODING APPARATUS, characterized in that the apparatus additionally comprises: a context modeler configured to determine a context enhancement parameter to determine a context index based solely on whether the transformation depth of an included transformation unit in a coding unit is equal to zero and obtain a context model for arithmetic coding of a significant coefficient indicator of the transformation unit of a luma component indicating whether there is a non-zero transformation coefficient in the transformation unit of the luma component using the context index obtained by adding the context boost parameter and a context shift; and an arithmetic encoder for arithmetic encoding the transform unit significant coefficient indicator based on the determined context model, wherein the transform depth indicates the number of times the encoding unit is divided so as to determine the transform unit included in the coding unit, where the transformation depth indicates the number of times the transformation unit is divided to determine the transformation unit included in the coding unit, where when the transformation depth is equal to zero, the modeler Context sets the context boost parameter to one; and when the transformation depth is not equal to zero, the context modeler sets the context boost parameter to zero, when the division transform indicator for the current transform unit of the current transform depth k indicates a division, a transformation unit of the current transformation depth k is divided into four transformation units with a lower transformation depth (k + 1). 3. COMPUTER-READABLE, NON-TRAVELING MEDIUM FOR STORING DATA ASSOCIATED WITH A VIDEO, characterized by comprising a bit stream stored on the non-transient computer readable medium, the bit stream including: a division transform indicator used to determine a unit of transformation to the inverse transformation process; and a luma component transformation unit significant coefficient indicator indicating whether there is a non-zero transformation coefficient in the luma component transformation unit, wherein the luma component transformation unit significant coefficient indicator is obtained by decoding Bitstream arithmetic based on a context model, the context model being determined by using a context augmentation parameter to determine a context index based only on whether a transformation unit transformation depth is included in a unit of encoding is equal to zero, and the context index being obtained by adding the context boost parameter and a context shift, where the depth of transformation indicates the number of times the encoding unit is divided to determine the unit of transformation included in the coding unit, where, when the depth of the transform is equal to zero, the context boost parameter is set to one; and when the transformation depth is not equal to zero, the context boost parameter is set to zero, when the division transform indicator for the current transform unit of the current transform depth k indicates a division, the transform unit of the current transformation depth k is divided into four transformation units with a lower transformation depth (k + 1).

Images